Statistics, Power, and Expertise

When we think of knowledge in the context of government, we often think of statistics. In fact, it’s arguable that statistics are not merely an especially prominent form of politically useful knowledge, but that their increasing use, starting in the seventeenth century and gathering pace ever since, was precisely what gave rise to our very idea of expertise in a political and administrative context — from number-crunching technocrats to low-level bean-counters — in the first place. So attitudes towards numbers and their use offer at the very least a useful point of entry in exploring the current assault on expertise in all its forms. They may even be at the heart of it.

Writing in The Guardian‘s Long Read recently, William Davies does a wonderful job of relating the history of statistics as a form of expert knowledge to the current crisis of confidence in expertise. Particularly valuable is his use of history to get beyond what he describes as a polarization of views, one side seeing statistics as “unquestionable truths” (putatively, the “expert” view), and the other regarding them as “elite conspiracies” perpetrated on ordinary people by political insiders, academics and bureaucrats (the populist take). Davies is surely right that statistics are neither of these things but are, rather, “tools designed to simplify the job of government, for better or worse.” He is also right to detail genuine limitations and serious critiques of statistics — and to distinguish these from the cheap, shotgun-style anti-intellectualism of our present moment. Blanket disdain for difficult ideas or unwelcome findings does not open expertise up for examination, but turns it instead into a black box best left closed.

Still, I couldn’t help noticing a couple of questionable points in Davies’s article respecting the origins of statistics in the seventeenth century. Full disclosure: I wrote a book on one of the key figures in the story, William Petty (1623-1687), and since there is no broad narrative in which a narrow specialist will not find holes, the risk of academic nit-picking here is high. The points that struck me are, in the sweep of Davies’s article, small. They should neither derogate from its value nor distract from its argument about the differences between statistics and “big data”. They might, however, cast the subject in a slightly different historical light.

Davies presents the emergence of statistics as the outcome of two convergent processes: the transformation of states in the context of war and civil strife (think: religious conflict, military technology, fiscal pressure), and the transformation of scientific theory and practice by empiricism and quantification (think: Bacon, Galileo, Descartes). While the former set of developments prompted governments to focus their attention on population as a source of manpower and revenue, the latter promised to represent social and economic facts with scientific accuracy and mathematical precision. Thus were statistics — a form of standardized, objective, holistic, and authoritative knowledge about the nation — and their purveyor, the expert, born: “Pioneering English demographers such as William Petty and John Graunt adapted mathematical techniques to estimate population changes,” Davies writes, “for which they were hired by Oliver Cromwell and Charles II.”

Yet although the first part of the sentence just quoted is true, the second is quite false. William Petty, MD, was hired by Cromwell’s regime as a physician to its army in Ireland, and he subsequently wheedled his way into running the land survey that followed the Cromwellian conquest of that country. Neither task involved, indeed both predated, his demographic work. He was later, briefly and disastrously, a judge of the Irish Court of Admiralty under Charles; his statistics did not figure there either. Although he tried assiduously through the 1670s and 1680s to promote “political arithmetic” as means of making policy using demographic data, directing dozens of manuscripts to powerful patrons, the longed-for job never came. His disappointment is captured in one of his many laments on the subject to his cousin and friend Robert Southwell:

I do not reckon all my right and reasoning to be worth a straw, till I can get some powerfull person to consider it. I compose curious peeces in Music and play them accurately; but all this while my hearers are as deaf as haddocks, nor will their eares (I feare) ever be opened with just applications, till Wee run a Spit into them made of some convenient Metall.[1]

The London tradesman John Graunt (1620-1674), meanwhile, published a single — albeit foundational — work of demographic analysis, the 1662 Natural and Political Observations… upon the Bills of Mortality, and was not heard from on the subject again. He was elected to the Royal Society for his pains, but he was never employed by either Cromwell or Charles.

Does it matter? One might be tempted to treat Davies’s account as a bit of unintentional foreshortening. Perhaps Petty and Graunt didn’t find state employment, we might say, but their work soon did, and surely that’s the key thing. But this not only misses the element of “contingency” that historians tend to emphasize; it also misses an important point about the advent of statistical expertise. Petty’s work was widely known in his own lifetime. The manuscripts that became his major works were read by powerful men. He was regarded as extremely if not uniquely knowledgeable on matters of population and of Ireland. But he was, in the end, not only not hired to modernize the state, but was in fact seen as almost unemployable — full of bold ideas but, as Charles II put it, forever “aiming at impossible things.” “Expert” did not then exist as an English noun. Instead, the contemporary term that fit Petty best, and which he tried hard to shake, was “projector“. And the projector’s knowledge was, in multiple senses, suspect.

This brings me to the second dubious point in Davies’s article: that what Petty offered was “a way of thinking about populations that privileged… objective facts.” Unlike the claim about Cromwell and Charles II, which is just inaccurate, this is a matter of interpretation. Davies’s starting point is that the present-day assault on expertise has thrown statistics from their position as neutral “reference points that everyone — no matter what their politics — can agree on”; it is therefore with the valuation and pursuit of “objective facts” in the age of the Scientific Revolution and the burgeoning bureaucratic state that his history of statistics naturally begins. This is a familiar reading of seventeenth-century quantification, and it lends itself to appealing narratives.

It’s also an interpretation that Petty embraced… sometimes. An avowed disciple of Francis Bacon, an avid experimental philosopher, and reputedly also a talented mathematician, Petty made a personal catchphrase of the notion that phenomena should be reduced to “Number, Weight, and Measure.” But consider, on the other hand, what he himself has to say about his numerical facts in his Political Arithmetick (circulating in manuscript by 1671, but printed only in 1690):

Now the Observations or Positions expressed by Number, Weight, and Measure, upon which I bottom the ensuing Discourses, are either true, or not apparently false, and which if they are not already true, certain, and evident, yet may be made so by the Sovereign Power, Nam id certum est quod certum reddi potest [For that is certain which may be made so], and if they are false, not so false as to destroy the argument they are brought for; but at worst are sufficient as Suppositions to shew the way to that Knowledge I aim at.[2]

Far from simply asserting their empirical accuracy, Petty suggests that his numbers are not so much “objective facts” as plausible estimates — or, at the very least, realistic possibilities. The facts are not so objective, then, as they might seem. To appeal to what is “apparent” is to appeal to a perceiving subject with a point of view that others may or may not share. Nor (and this is if anything even more important) are the facts neutral, if by neutral we mean apolitical in any straightforward sense. Because rather than the objective truth of his numbers guaranteeing their usefulness for policy, Petty here indicates that it is the other way around: it is the power of the sovereign that makes his numbers true. And indeed, for his hero Francis Bacon, “the discovery of products and results is like a warranty or guarantee of the truth of a philosophy.”[3]

To rephrase the above points, then, one might say, first, that statistics was not simply a transparently useful approach to universally recognizable facts eagerly grasped by the modern state and inevitably taken up by an increasingly inquisitive, modern public. Producing and publishing numbers as well as disseminating the interpretive framework within which they counted as facts — the point of view from which they were “not apparently false”, if not “already true” — was the subject of an immense, and often unsuccessful, proselytizing effort. Second, this effort was aimed at the state not just, as Davies suggests, because only the state could use or gather the kind of national numbers involved, but also because the power of the state to transform (not just measure) populations — to make the numbers true, to create the reality they described — was what justified the enterprise in the first place. The quantification of population was a “project” both in the banal sense that it took targeted effort to achieve and in the seventeenth-century sense that it was designed to transform the world — and, in doing so, to serve a combination of public and private interests. It was political through and through.

This becomes obvious if we consider the specific populations and policies at issue. For Petty this meant, first and foremost, the Irish; quantification was a means to transform, or “transmute them into English”, to use his words, though a program of forced migration and intermarriage between socially marginal English women and Irish men. Later he turned his work to the demographic balance between Catholics and Protestants in all three Stuart Kingdoms, proposing massive “transplantations” of people — in effect, a kind of transnational gerrymandering scheme — to create populations that would adhere to a Catholicizing monarchy under Charles’s brother and successor James II. Petty’s own successors used “political arithmetic” to help bring about the Union with Scotland; to develop schemes and institutions for governing (and “improving”) criminals, orphans, the poor, and the colonized — and, of course, to facilitate Britain’s increasingly central role in the transatlantic slave trade. That seventeenth- and eighteenth-century political arithmeticians, and later statisticians, also investigated population trends over time and relationships between demographic processes and various environments, creating new discourses of risk and probability, life expectancy and public health, is important. But these apparently more recondite or respectable pursuits were neither fundamentally separate from their political context nor different — either in kind or in origin — from the statistical practices used in the more egregious essays in social engineering to which those contexts also gave rise. Nor did the embedment of statistics in politics end in the modern period, whether we think that began in 1789, 1815, or 1914. Google “statistics” and “eugenics” if you don’t believe me.

All of which is to say that looking a little more carefully at the seventeenth-century origins of statistics shows us some of the ways in which expertise as an idea, and statistics as a technology, have been tainted by interest and politics, indeed by a kind of violence, from the start. Petty himself was not merely mocked (as he perhaps thought) for being ahead of his time, but also criticized and kept from office for his very real conflicts of interest in Ireland, where he was both director of Cromwell’s survey and one of the major beneficiaries of the resulting land settlement. Accusations of personal gain attached easily to “projectors” peddling shiny, new ideas to the ramshackle early modern state; they were like nothing so much as alchemists in this regard, and indeed there was social and intellectual overlap between the two groups. It was often from just such outsiders — “enthusiastic amateurs”, in Davies’s kind rendering — that the most ambitious schemes for extending the reach of the state or of potential patrons within it came. It may be that as projectors became experts and projects became disciplinary and bureaucratic practices, suspicions of private interest and scepticism about grand promises metastasized, in some quarters, into a distrust of experts, expertise, and the kind of government they made.

By the foregoing I do not mean to embrace any such wholesale dismissal. As a historian, of course, I have a personal and professional stake in the idea that knowledge is possible through particular kinds of work, and that these can be meaningfully assessed as well as taught, even if they have substantial subjective elements in practice. That aside, however, I also think it perfectly possible to understand that knowledge, including science, is made in and marked by its social and political contexts without abandoning the view that some ideas are truer, more accurate, more reliable than others. In fact, it’s vital to hold onto both ideas. To contrast an idea of straightforwardly neutral, objective science with a sudden upsurge of demotic rejection leaves some important parts of the story, to my mind, out of view. On the other hand, to point out that an expert has a point of view is to establish not that his or her ideas are worthless but merely that he or she is human and exists in history. As we all are and do.


[1] William Petty to Robert Southwell, 2 August 1681, in Marquis of Lansdowne (ed.), The Petty-Southwell Correspondence (London, 1928), 91.
[2] William Petty, Political Arithmetick (London, 1690), sig. A4v.
[3] Francis Bacon (trans. Lisa Jardine and Michael Silverthorne), The New Organon (Cambridge, 2000), 60.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s